The Final Theory of Physics - a Tautology?

C. Baumgarten1
5244 Birrhard, Switzerlanda)
(Dated: 7 February 2017)
We acuminate the idea of a final theory of physics in order to analyze its logical implications and consequences.
It is argued that the rationale of a final theory is the principle of sufficient reason. This implies that a
final theory of physics, presumed such a theory is possible, does not allow to incorporate substantial (non-
trivial) propositions unless they are logically or mathematically deduced. Differences between physics and
mathematics are discussed with emphasis on the role of physical constants. It is shown that it is logically
arXiv:1702.00301v2 [physics.gen-ph] 6 Feb 2017

impossible to introduce constants on the fundamental level of a final theory. The most fundamental constants
emerging within a final theory are constants of motion.
It is argued that the only possibility to formulate a final theory is necessarily a tautology: A final theory of
physics can only be derived from those presumptions about reality that are inherent in the idea and practice
of physics itself. It is argued that a final theory is based on the notion of objectivity, but it is logically
impossible that an ideal final theory supports realism.

PACS numbers: 01.70.+w,02.10.-v,03.65.Ta,03.65.Ca

Keywords: Philosophy of science, Logic, Measurement Theory, Formalism of Quantum mechanics

I. INTRODUCTION known fundamental particles and forces, it neither im-

In his book Dreams of a Final Theory Nobel laureate
Steven Weinberg wrote1 : “It is difficult to imagine that
we could ever be in possessions of final physical principles
that have no explanation in terms of deeper principles.”
But Weinberg (of course) did exactly this: he described
the idea of a final theory in his book. Not all physicists,
not even all Nobel laureates, share Weinbergs expecta-
tion that such a theory might be possible2 . Nevertheless
we shall pick up and discuss various aspects of Wein-
bergs dream. Specifically we shall use the idea of the
final theory to discuss the relation between physics and
mathematics. Whether we believe or not that a final is
possible in principle, both cases have their specific im-
plications 1 . Weinberg’s dream concerns a theory that is
final because no deeper theory is possible. The concept of
a final theory (FiT ) understood in this way differs con-
siderably from that of a grand unified theory (GUT ) or
a theory of everything. A GUT aims to provide a unified
description of all fundamental forces of nature. However
revolutionary this would be, it remains within the scope
of already existing theories; the concept of a GUT does
not necessarily imply finality, but completeness with re-
spect to known experimental facts.
The conjectured finality results, as Weinberg explains,
from “the beauty of simplicity and inevitability, the
beauty of perfect structure, the beauty of everything
fitting together, of nothing being changeable, of logical
rigidity.” The keyword here is inevitability: Though the
idea of a GUT implies completeness with respect to the
plies simplicity, nor a deep or even final level of under-
standing, nor a high degree of explanatory power. A
GUT could be ugly and incomprehensible. The form of
a GUT - if we extrapolate the prominent tendencies in
theoretical physics - will likely consist of a set of mathe-
matical structures that allow to summarize the structure
of the standard model of particle physics. But the very
idea of a GUT does not include explanations why na-
ture should prefer a specific mathematical structure and
not some other. Or, in other words, a GUT would -
in contrast to the ideal FiT - not exclude the possibil-
ity that different universes with different laws of nature
might exist at least in principle.
However, we shall argue that the conjectured finality
of a FiT follows a different logic than the assumed com-
pleteness of the GUT . Furthermore we shall argue that
and why Weinberg’s specifications can, if at all, only be
met by a theory that is based on trivial premises. The
finality and inevitability that Weinberg expects from a fi-
nal theory implies that it is similar to pure mathematics.
This has a number of consequences 2 .
Weinberg wrote “But (I repeat and not for the last
time) I am concerned here not so much with what sci-
entists do [...] as I am with the logical order built into
nature itself”. We we shall argue that a final theory (if
possible at all) will be based exactly on this: on what
physicist do, what they have to do, in order to formulate
a quantitative and objective description of nature.
In this essay we further elaborate the premises that
we used in two preceeding publications20,21 in which the

2 One implication of the inevitability seems to be that the final the-

a) Electronic mail: christian-baumgarten@gmx.net ory is incompatible with the “Mathematical Universe Hypothe-
1 It might be the historical fate of physics to fail providing a final sis” in Tegmark’s sense of enumerable possible universes5 : If
theory, it might end up with a last instead of a final theory - but many worlds are possible, then none of them can be regarded as
we will not consider this possibility. inevitable.

author sketched a logical line of reasoning that allows to
derive relativity, electrodynamics and the Dirac equation
from (almost) trivial principles. Here we shall argue that
these principles are a promising starting point of a final
theory - not because we are able to deduce a GUT from
it, but because they scetch a possibility to achieve the
required logic of a FiT .
We shall argue that specifically the inevitability and
finality of a physical theory can only be derived from a
systematic analysis of the form or method of physics itself:
Since the theory has to demonstrate that any physical
world inevitably has the form of our universe, then it
can not be purely empirical: it has to demonstrate logical
necessity. And since we do (like Weinberg) not believe
that it can be derived “from pure thought”, then it can
only be derived as an tautology, i.e. from a description
of physics itself.
In Sec. II we will try to precisely specify what we con-
sider to be a final theory of physics. We do not know if
our definition is in full agreement with Weinbergs idea
of a FiT and his attitude towards such a theory. We
will acuminate the idea to an extend that allows to draw
conclusions. Insofar it is a thought’s experiment.
In Sec. III we discuss aspects of the relation and the dif-
ferences between mathematics and physics and the conse-
quences for a FiT . Here we put special emphasis on the
difference in the methods of concept formation between
mathematics and physics and what can be concluded for
a final theory.
In Sec. III A we investigate the role played by the so-
called “physical constants”, both, dimensional and di-
mensionless. We shall discuss where and how such con-
stants may be introduced or derived within a final the-
ory. The specific meaning of natural units is considered in
Sec. III B, their role within physical laws in Sec. III C and
their importance with respect to the concept of objectiv-
ity in a final theory, in Sec. III D. In Sec. IV we draw first
conclusions and discuss their implications, mainly that
the foundation of a final theory can only be formulated
as a tautology. In Sec. IV A we describe the tautology,
namely we provide a simple description of the premises
of physics.
In Sec. IV B we scetch some consequences concerning
the attitude of science towards reality. We shall argue
that the formulation of a FiT first of all requires to with-
draw all metaphysical presuppositions. In Secs. IV C,V
we scetch the conclusions for the form of a final theory.
In Sec. VI we shall argue that there is an intrinsic incom-
patibility of rationalism with realism.
Finally we shall summarize our considerations in
Sec. VIII.
II. THE SPECIFICATIONS OF A FINAL THEORY
The current idea of a GUT is that of a theory which
unifies all forces of nature into a common conceptional
framework. We do not intend to discuss the various can-
2
didates on the market like quantum gravity or the various
string theories. It is sufficient to notice that these theo-
ries aim to be unified theories. But a final theory is more
than just a unification, it is a theory which leaves no ques-
tions open. It not only allows to derive all known types
of particles and interactions, but also the dimension of
space-time. But even this is not sufficient to provide fi-
nality: As we shall argue, the finality implies that it even
provides a formal derivation of the fundamental concepts
of physics.
This specification might seem exaggerated and the idea
of the possibility of a FiT hubristic, however this is not
necessarily the case. The idea of a FiT only exemplifies
the ratio behind the scientific project itself: Weinberg’s
dream is a consequence of the well-known “principle of
sufficient reason” (PSR) and it claims nothing more than
that everything in nature should have a reason7 . Indeed
it is difficult to imagine and to argue that some fact,
or law, or relation in nature might have no reason at all,
that it might be based on pure contingency, that the mass
ratio of proton and electron is unexplainable. It would
simply be unreasonable and alien to science to claim that
physical relations or some law might have no reason. Be-
cause - if something in nature has no reason, why should
anything have reason? Hence science has to presume rea-
son, and that is why it is reasonable to consider that a
FiT could be within reach some day - since anything that
is reasonable can, by definition, be understood 3 . Tech-
nical or financial limitations might prevent physics from
performing the required experiments, or the mathemat-
ics might become so difficult, that we will never be able
to solve all riddles of physics. But to question the PSR
would go beyond that - to some degree it would mean to
question the scientific project itself.
Gerard t’Hooft expressed his attitude towards the idea
of a FiT as follows2 : “[...] in particular string theorists
expect that the ultimate laws of physics will contain a
kind of logic that is even more mysterious and alien than
that of quantum mechanics. I, however, will only be con-
tent if the logic is found to be completely straightfor-
ward.” Or Leonhard Susskind (ibd.): “Not only would
the theory explain why the proton and neutron are about
1,800 times heavier than the electron, but the theory
would also explain itself: no other theory is possible.”
If we include these requirements in the specifications of
a FiT , then the theory has to carry the evidence for its
truth in itself. This means that a FiT has to demon-
strate that the laws governing physical reality could by
no means be different than those formulated by the FiT
. Then of course the FiT has to provide a formal cer-
tainty and “logical rigidity” as otherwise only known in
mathematics.
Stanley Goldberg characterized the logic of mathemat-
ics as follows18 : “Different branches of mathematics have
3
This does not imply that everything must have an intelligible
cause, though.

different rules but in all branches, since the rules are pre-
determined, the conclusion is actually a restatement, in
a new form, of the premises. Mathematics, like all formal
logic, is tautological. That is not to say that it is uninter-
esting or that it doesn’t contain many surprises.“ Hence
a FiT might be possible if it is formally equivalent to
mathematics. This raises the question of the differences
between mathematics and physics and if and how they
could be overcome. We shall come back to this in Sec. III.
The required self-evidence of a final theory (considered
that such a theory is possible) does not allow to base
the theory on substantial propositions, since substantial
propositions are not self-evident. Using Goldbergs phrase
we might say that substantial propositions are those that
are not predetermined. All propositions that can be de-
rived within the FiT must be inherent in some form in
the premises of physics or they must be derivable math-
ematically. To substantiate the meaning of this idea, let
us consider some examples of what propositions might
be regarded as substantial. It is a substantial claim that
space-time must have 3 + 1 dimensions and to believe
that space is fundamental. Though it is clear that our
world appears to have this form, it is not evident why this
should be the case - why any physical world should have
3 + 1 dimensions. One might have doubts that physics
will ever be able to give clear and distinct reasons why
any possible physical world must have 3 + 1 dimensions,
but one can not deny that the question “why 3+1 dimen-
sions?” itself is legitimate and meaningful. Therefore a
statement that the physical world has (or must have)
3+1 dimensions, is substantial and requires a derivation
within a FiT . Hence, if we believe that a FiT should
be possible and that it has to provide such explanations,
then any physical world must have 3 + 1 dimensions.
A GUT might be based on substantial assumptions
and postulates, for instance a specific dimensionality of
space-time or some special general group structure or on
the postulation of some type of string: A GUT could be
formulated as a theory of postulates. However, Einstein
remarked that “When we say that we understand a group
of natural phenomena, we mean that we have found a
constructive theory which embraces them”3 . Following
Einstein, a GUT might as well be a theory of principle,
but a FiT has to be anticipated as a purely constructive
theory: It has to explain the 3+1 dimensional space-time
by construction, it has to explain, why and what physi-
cal significance specific groups have, why there must be
a maximal velocity of massive objects, why Maxwell’s
equations have their specific form and so on. This im-
plies that a FiT (if possible at all) also has to reformu-
late relativity; it has to replace all of its postulates by a
reasonable construction. The same would hold for quan-
tum mechanics. A final theory has to explain why we
(have to?) use complex numbers in quantum mechanics
and why the momentum should be equal to the spatial
derivative of the wave function (times the unit imagi-
nary). A FiT is a deductive theory in which all elements
are deduced from unquestionable first trivial principles.
3
III. WHAT DISTINGUISHES PHYSICS FROM
MATHEMATICS?
Since the final theory is constructive, it can not sim-
ply presume concepts like mass or velocity that originate
from experience. The required inevitability implies that a
final theory has to demonstrate how quantities like mass,
energy, charge and field emerge within a mathematical
framework that is developed on the basis of self-evident
abstract principles. However there is at least one con-
cept that has to be presumed: Time. The reason (and
legitimization) is the inevitability of this concept. There
is not way to describe the meaning of physics without
presuming the notion of time in some way.
Goldberg described this in his book on relativity18 as
follows: “The undefined terms for the dimensional sys-
tem we will be using are ”length,” ”time” and ”mass.”
These are the common, undefined primitives of physics.
Although I can tell you how to measure a length, or a
time, or a mass by a formula, the concepts themselves are
undefined. Either you know or you don’t know what I
mean when I use a phrase like ”time passes.” The dimen-
sions of all other quantities in physics can be expressed in
terms of these fundamental quantities.” This means that
the “understanding” of certain concepts can only be ob-
tained by personal experience. If all substantial concepts
of physics have to emerge within the final theory then
a FiT is formally equivalent to mathematics. And then
the “mapping” between mathematical form and physical
meaning is reversed with respect to the usual method of
physics. If substantial physical concepts like “mass” and
“energy” or “spin” can not be presumed, these notions
have to be identified from pure mathematical form.
Usually the arrow that describes the formation of con-
cept in physics, points from the phenomena to the equa-
tions. In mathematics one also finds the opposite di-
rection: Abstract concepts that have been invented on
the basis of pure formal considerations and then obtain
a name from conventional language like “fiber” or “pen-
cil”. To establish a relationship between the mathemat-
ical form and its physical content is called interpreta-
tion. And in contrast to mathematics, the interpreta-
tion charges variables with physical meaning: the math
of physics refers to things outside mathematics. It would
be meaningless to ask whether names given to purely
mathematical objects or functions are “correct”. Math-
ematical objects do not refer to anything else outside
mathematics. But a FiT has to identify mathematical
structures with physical objects. There is no formal way
to prove the truth of this mapping. We can only verify
(or falsify) its experimental predictions. But the vari-
ables that correspond to these predictions are a subset
of all variables of the model: We measure mean values
- even in the microscopic case. Phenomenological phys-
ical models that describe specific physical problems are
to a large degree self-sufficient. It is their purpose to
provide predictions. But a FiT has a different purpose,
namely to provide the bridge between phenomenological

theories. It must be complete and convincing, though it
is questionable, whether the required interpretations can
be unique: the effectiveness of physical models is that
they are applicable to a variety of phenomena.
If the phenomena (the meaning) has not been estab-
lished within physics beforehand or if the FiT would de-
scribe completely unknown phenomena of nature, then
the attempt of interpretation might be a serious chal-
lenge: The meaning of variables can (safely) only be ob-
tained from already known physical concepts. That is,
the path of reasoning within a FiT can not start at the
phenomena and end in equations, but vice versa. The
equations have to be derived on the basis of the abstract
principles and the resulting variables in these equations
have to be identified with physical quantities. This is a
task of pattern recognition or - as Walter Smilga named
it - “reverse engineering”6.
This is a characteristic difference between mathemat-
ics and the classical method of physics: Mathematicians
define or derive abstract objects which are invented in a
purely formal way, as a kind of game, a game of thought.
Then, in a second step, the forms obtain their names,
which are essentially arbitrary. Physics however picks
properties known from everyday experience on the basis
of common language. These properties have to be defined
scientifically but this should not alter their meaning, it
should just render their meaning scientifically, i.e. quan-
titatively. The meaning of the “mass” of a material body
is essentially the same in physics and in common sense,
or at least it is very close.
When the physical significance of some symmetries
have been understood, then physical theory might pre-
dict new particles because the specific symmetry prin-
ciple suggests the particle for its completeness. But to
invent a new symmetry principle without any prior ex-
perimental knowledge can be regarded as almost impos-
sible. Hence the FiT requires for its formulation that
most pieces of the final puzzle are already known 4 .
Elementary particle physics is a special case in this
respect and its methods are closer to those of mathemat-
ics: Most properties (or components) of particles are ab-
stract and have no counterpart in the macroscopic world,
like for instance (iso-) spin or parity or helicity. In this
case physicists have to invent new, essentially arbitrary,
names for particle substructures (“quarks”) and proper-
ties (“strangeness”, “charme”). The deeper physics dives
into the micro-cosmos, the closer it seems to be to pure
4 Neither Leibniz nor Platon nor Newton could have derived a
final theory from pure thought, simply because the phenomena
to be explained where not known precisely enough. Platons list
of solids insofar proves the impressive power of pure thought as
Platon obviously understood the importance of symmetry prin-
ciples. But Platon could not have arrived by any means at the
gauge group SU (3) × SU (2) × U (1), not even if the mathematics
would have been available: Without the knowledge of the phe-
nomena, i.e. the list of particles, there would be no indication of
its physical significance.
4
mathematics. Without the knowledge of the empirically
developed physical concepts and notions it is hardly pos-
sible to recognize the physical meaning of the variables
appearing in the supposed final theory and to find inter-
pretations that provide a physically meaningful theory
from purely formal (mathematical) relations. If a FiT is
able to deduce equations from abstract principles, then
the “correct” interpretation is not included for free.
Physics and mathematics require certain notational
conventions, a nomenclature, that assigns symbols to cer-
tain quantities and relations. For instance, a2 + b2 =
c2 is associated with the Pythagorean theorem, while
m2 + p2 = E 2 , though of identical form, reminds us
of the relativistic energy formula in units where c = 1.
A purely formal derivation of equations within a FiT
as suggested by an analogy with pure mathematics does
neither provide the correct symbols nor the correct inter-
pretation: it only provides mathematical relationships,
the form. If concepts like “mass” and “energy” are sup-
posed to emerge from the formalism then, at some stage,
the FiT must provide an equation that matches, for in-
stance, the mentioned form of the relativistic energy for-
mula. However the FiT can not provide a proof that
the derived formula really means the relativistic energy
formula. A formally derived FiT does not automatically
provide physical but rather mathematical relationships.
The form provides information only in the eye of the
well-prepared observer. The transformation of a mathe-
matical form into a physical meaningful law requires an
identification of the terms appearing in equations with
physical quantities. This process is to some degree heuris-
tic: It is an interpretation. Therefore it is unlikely that
equations that emerge within a final theory can be in-
terpreted correctly by pure thought. The formulation
of the correct principles alone is not sufficient to obtain
physical meaning: Physical laws have to be known and
proven meaningful in advance, before a FiT can be for-
mulated. Modifying Goldberg’s insight, we might say
that we already have to know the form and relevance of
the Lorentz group in order to recognize its significance
within the mathematical framework of a FiT . Therefore
it is unlikely that a FiT could be formulated “from pure
thought” before physics sufficiently progressed towards a
fundamental level of physical reality: We can only rec-
ognize known patterns; the pure form does not generate
meaning. When the abstract inevitable principles of a
FiT are found, and a mathematical formalism has been
developed then one still has to establish the mapping be-
tween an abstract mathematical form and known physi-
cal laws. This step in the formulation of the FiT might
be summarized as “function follows form”. Though the
principles of the FiT must be trivial (“self-evident”), this
does not necessarily hold for the formulation of the the-
ory. Therefore it seems unlikely that a FiT will - from
beginning on - be recognized as a possible candidate for a
theory of everything. This seems to be sufficiently proven
by contemporary candidates like string theory: The ini-
tial ideas of theories might be simple, but the resulting

structures and mathematical forms must (by definition)
be rich and complicated enough to allow for the descrip-
tion of a large number of particles and fields. But if they
are rich and complicated enough, then they are difficult
to interpret.
The basic system of “undefined” quantities, that Gold-
berg mentioned, is still debated, but it appears obvious
that certain quantities can not be defined by other means
than by a description of their respective measurement
method or directly by reference to experience. This is
for instance the case for the notion of time. We can not
“explain” time to someone who does not know what it
is because it is unique. A unique concept can not be ex-
plained by other similar concepts since it is unique. But
though clocks don’t explain or define what time is, clock’s
define how we measure time. And this is all that is re-
quired: Physics is neither able nor obliged to tell what
things are, but physics can tell us how things behave.
It could be argued that, from the complete knowledge
of how things behave, we finally obtain an idea of what
things are. At least one might finally say, “A” and “B”
behave as if they were “X” and “Y”. The quantum me-
chanical wavefunction behaves as if it was a probability
density, but we can’t tell seriously that it really is (and
nothing else). And it is not even obvious what difference
it makes.
The impossibility to define the concept of time is in-
sofar unproblematic in this context as the concept of
physics itself presupposes time: The formulation of the
“laws of nature” is meaningful because laws allow to pre-
dict measurement results. Physics is and always was
more than an intellectual excercise. It is a science that
aims to provide new technologies, new methods that al-
low to manipulate objects.
A physical prediction is in several ways related to time:
First of all, predictions enables to know the value of a
measurable quantity before the measurement has been
done, or, as in quantum mechanics, it at least enables to
predict the possible measurement results and their proba-
bilities. And secondly there is an indirect relationship by
the presumptions that are required to make a measure-
ment meaningful, namely, that we have reference objects
with constant properties (a clock, the prototype meter,
etc) and variable properties that are to be predicted. No
measurement without a reference. Hence the form of
physics implies and presupposes that there is time, i.e.
that the same type of quantity is allowed to appear in
two forms, as a variable and as a constant reference.
Therefore we shall next discuss the role of units and of
so-called “physical constants” in Sec. III A. Inherent in
the notion of time is also the concept of constancy and the
specific role of constants is an important difference be-
tween physics and mathematics. Some of these constants
are called “physical constants” because they do not ap-
pear in mathematics. These are first of all the constants
with dimension, i.e. those that have units, like the speed
of light c or the Planck’s constant h̄. Then there are
dimensionless constants like the fine-structure constant.
5
The latter could, at least in principle, be the result of
mathematical relations. A final theory has to provide
the values of these dimensionless constants - which im-
plies that they can be derived.
A. Measurement Standards And Physical Constants
Physics used arbitrary measurement standards in its
history. There is the tale about Galileo, who measured
the frequency of swinging oil lamps in church using his
pulse as frequency standard12 . The use of the own body
as a basic standard is a natural and obvious thing to do,
as the size of an object relative to ourselves is exactly
what we naturally want to know. The number of steps
on the way between point A to point B is exactly the
type of practical knowledge that one may assume as the
origin of measurement. And quite naturally, distances
can be and likely have always been measured in units of
time: A to point B might be a three day walk or a single
day horse ride. The use of the own body as a reference to
measure length is nearby, because the scale of our body
defines the scale of our world.
It is well within proportion that the meter is roughly
the size of a footstep. The unit of a foot testifies this
finding and shows that measurements started with scien-
tifically arbitrary units - standards that are defined prag-
matically. The first modern scientists like Galileo picked
phenomenological properties like length or duration of
something in motion and measured dynamical properties
of swinging pendula and falling stones. Consequently
the first “laws” are mere (idealized) descriptions of pro-
portions13 . First scientific observations were of the sort,
that the frequency of a swinging pendulum does (in first
approximation) not depend on its amplitude. This con-
tinued with Keplers and Newton’s laws: Physical laws
were rarely formulated as equations, but as statements of
proportionality. It should be noted that laws of propor-
tionality allow to circumvent the definition of universal
units. If we say that the frequency of a swinging pen-
dulum is inversely proportional to the square root of its
length, then the experimental confirmation requires (if at
all) rulers and clocks. The law itself is correct no matter
what units are choosen and hence a universal system of
units is not required to confirm them experimentally. De-
spite the fact that it is possible to formulate objectively
correct laws of proportion without explicit reference to
specific objects as measurement standards, the measure-
ments that are used to find these laws, of course require
rulers and clocks. But they do not need a universal gauge.
As technology evolved it became unavoidable to define
systems of units: laws of proportionality do not suffice for
the construction of a pendulum with a specified period.
A consequence of the formulation of the laws of nature by
the use of equations instead of mere laws of proportions
is the appearance of conversion factors like the Avogadro
constant or the gravitational constant. Often these con-
stants are named physical constants. Wikipedia explains

the term physical constant as follows: “A fundamental
physical constant is a physical quantity that is gener-
ally believed to be both universal in nature and constant
in time.”14 . This, however, does not reflect the under-
standing of the physical significance of physical constants
in theoretical physics. Dirac wrote: “The information
that experimentalists obtain provides us with a number
of constants. These constants usually have dimensions,
and then of course, they depend on what units one uses,
whether centimetres or inches. Then they are not of
theoretical interest.”15 . Dimensionful physical constants
bear no physical significance. This point of view has not
changed significantly until today16 - even the idea that
Planck’s constant is physically significant was questioned
with some convincing arguments17 .
This situation might appear paradoxical: Those physi-
cal constants that can expected to be the result of purely
mathematical considerations, are supposed to be those
with physical significance and a FiT would eventually
have to provide their values, while those with a physical
dimension, are considered to be physically meaningless.
But though the numerical values of dimensionful physi-
cal constants are meaningless, as Dirac explained, their
constancy is not. For instance, the numerical value of
the speed of light is meaningless as it depends on the
choice of units. But the fact that a maximal velocity for
massive objects exists, is of course meaningful, it is even
of fundamental importance. However, since a FiT must
be formally derived from abstract principles, the values
of dimensionful constants can not be the result of such a
theory. And since their values are not significant, there is
no reason why these constants should emerge from fun-
damental equations. In theoretical physics, the system
of units is often chosen in a way that the dimensionful
constants like c and h̄ can be replaced by unity, i.e. they
do not appear in the equations and we can expect the
same for a FiT .
For instance, it can be shown that the group velocity
vgr of wave-packets is given by vgr = ∇k ω(k). In quan-
tum mechanics (QM) and quantum field theory (QFT),
wave packets represent particles and therefore the veloc-
ity of these particles is derived from the respective disper-
sion relation. For massless particles this yields ω(k) = |k|
such that |vgr | = 1. If frequency and wave vector have the
same unit, there is no theoretical need for a dimensional
constant called “speed of light”. This means that a FiT
constructed from formal principles can not provide equa-
tions in which the so-called “physical constants” c, h̄, ε0
automatically appear. In a FiT such constants have to
be introduced artificially simply because these constants
are artificial.
B. The Significance of Natural Units
Dimensional physical constants loose their contingent
status only if natural units like the elementary charge
e or the electron mass me can be found which allow to
6
replace the contingent historical units with “objective”
units, i.e. the corresponding constant properties of fun-
damental objects 5 . One of the first known natural units
was the elementary charge e and it is indeed remarkable
that any amount of charge in nature can be expressed by
an integer number. If quarks are considered, the natural
unit charge is e/3, but still any natural amount of charge
is quantized and can be expressed as in the natural unit
represented by some fundamental object that serves as
reference.
Natural units refer to the properties of fundamental
(stable) objects (or processes) of nature. If matter would
not be composed of atoms, e.g. if matter would not be
quantized in some way, such fundamental objects could
not be explained within a FiT . It was a common idea
in the history of physics that matter would be something
that completely and continuously fills up space, pretty
much as it appears to the eye. This idea is nearby, directly
derived from the “empirical evidence”. Today we know
that the nucleus is very small compared to the atom’s
dimensions, which are defined by the electronic orbitals,
so that matter is far from filling up space.
However, if matter would really be a continuous some-
thing that fills up space-time in a way that could not
be derived from a more fundamental theory, then a
FiT would be impossible, because this would imply that
space-time is fundamental and hence unexplainable. It
has been suggested that the antropic principle might
serve as an explanation. But it hardly explains anything;
it does not generate the kind of reason that the principle
of sufficient reason recomments.
Without a description of the internal dynamics of such
a material, we could by no means argue that and why
a specific radius should appear. The only way to obtain
a granular structure of matter would then be to postu-
late some kind of quantization. Otherwise there would
be no means to replace the arbitrary units of mass and
charge by natural units and a final theory in the sense of
Weinberg would be impossible. Without quantization we
could not refer to the respective properties of an electron
and it seems clear, that a FiT would be unthinkable in
such a world as there would be no way to get rid of the
arbitrariness of dimensionful units. This implies that a
classical FiT (i.e. a FiT without some type of quantiza-
tion) is logically impossible. And vice versa: If we spec-
ulate that the formulation of a final theory is possible,
then the FiT has to contain quantization in some way.
Otherwise we could not derive the objects that provide
the required natural units.
5 We emphasize the importance of reference standards within the
logic of physics: objectivity is based on the use of reference ob-
jects. Let us expand the notion of objects and include properties
of reference processes, despite the fact that we will later conjec-
ture that objects can - within a final theory - only be understood
as processes.

C. The Form of Physical Equations
Given two physical quantities E and p of different di-
mension are supposed to be related by a physical law in
the form
E = f (p) . (1)
Maybe we derived this law by a sequence of measure-
ments Ek in dependence of pk . The range within which
one can experimentally prepare p and measure E is cer-
tainly finite, such that one can always approximate f (p)
by a polynomial of finite degree. In the simplest case,
we find that f (p) is proportional to p or p2 . If f (p) is a
steady function, this has to be expected for small enough
ranges of p, as one may always write a continuous func-
tion f (p) of a continuous variable p as a Taylor series:
7
If the values of these constants would be independent
such that we could not find a mathematical law that al-
lows for their derivation, then such a law needed an in-
finite number of dimensionless physical constants that
would have to be the result of a FiT . But since the FiT
must itself be analogue to mathematics, this would be a
contradiction in itself. As we specified that a FiT must
derive all dimensionless physical constants mathemati-
cally, then the most fundamental laws may not contain
meaningful dimensionless physical constants. The idea of
a FiT implies that the dimensionless physical constants
are effectively mathematical constants. Hence the Tay-
lor series is a short polynomial or it is a transcendental
function. Since the most fundamental equations of a FiT
may not include dimensionless constants, these most fun-
damental laws have to be simple. Not because physicist
prefer simplicity, but because the laws are assumed to be
∞  fundamental. Nature might turn out to be different, but
P df (k) (p)  (p−p̃)k
f (p) = dp  k!
k=0 p̃
(2)
= C0 + C1 (p − p̃) + C2 (p − p̃)2 + · · ·
However, we can only add physical quantities of the same
unit. Since p has a dimension, then all coefficients Ck
have different dimensions, C0 the dimension of E, de-
noted by [E], C1 has dimension [E]/[p], C2 the dimension
[E]/[p2 ] and so on. This would lead to an infinite number
of dimensional constants. However, as only two dimen-
sionful quantities are related in Eq. 1, it is sufficient to
establish two dimensionful non-zero constants p0 and E0
and to express f in the form
E = E0 f (p/p0 ) , (3)
where now f (x) can have any mathematical form since
x = p/p0 is dimensionless. These two dimensionful con-
stants would fall into one, if the Taylor series had just a
single term, i.e. if the true law would be
E = E0 (p/p0 )n , (4)
where we assume n 6= 0 to obtain a non-trivial equation.
However, in this case, we could not determine two con-
stants p0 and E0 separately, but only the proportionality
constants C = E0 /pn0 .
This example demonstrates that the mathematical
form of physical laws and the number of physical con-
stants are indeed related. If a physical relation is contin-
uous and can be written as a Taylor series (Eq. 2) and
the required number of natural units E0 and p0 can be
found, then one can rewrite Eq. 2 in the form
x = p/p0
if it is, then the specified FiT is logically impossible.
Let us assume the true law would be
f (p) = E0 (exp (p/p0 ) − 1) , (6)
If p0 is very large compared to the experimentally real-
izeable range of p, then the first experimentally verified
law might be
f (p) = C1 p . (7)
This law would neither be absolutely right nor absolutely
wrong. It would be adequate within a certain range of p
and increasingly inaccurate and finally useless, if the p of
interest is not within or not even close to the mentioned
range. It is possible to determine C1 by the measure-
ments within a limited range of p, but not E0 and p0
separately. With progress of the experimental technol-
ogy one might be able to extend the range of p towards
and beyond p0 , such that we would be able to find
 2 !
p 1 p
f (p) = E0 + , (8)
p0 2 p0
and maybe with further experiments (or with a deeper
theoretical insight) we might conclude that the true law
would have to be Eq. 6. It is obvious that the first dimen-
sionful constant C1 has, as such, no fundamental mean-
ing. But what about E0 and p0 ?
Such constants have fundamental physical significance
given that they represent natural units E0 and p0 or that
they can be expressed by other natural constants. In the
relativistic energy-momentum relation
p
E = m2 c4 + p 2 c 2 (9)
∞ 
P dE (k) (x)  (x−x̃)k these two constants are m and c, such that with E0 =
E(x) = E0 dx k! (5)

k=0 p̃ m c2 and p0 = m c one may write instead of Eq. 9
= E0 (1 + D1 (x − x̃) + D2 (x − x̃)2 + · · ·) p
E/E0 = 1 + (p/p0 )2 . (10)
with an (potentially) infinite number of dimensionless The mass m is then a property of the electron (for in-
constants Dk . stance).

As a final theory may not contain arbitrary units but
only natural units as references, these units play a fun-
damental role for the way, in which physical laws are
and can be formulated within a FiT . The natural units
however are properties of reference objects. The FiT can
not derive the value of the mass of a reference object:
The FiT expresses the mass of other objects in units of
the reference object. The FiT can also not explain the
stability of the reference object: The FiT instead uses
the stability of the object as a self-sufficient fact: In an
objective reality we must have stable objects.
However, the FiT may not presume dimensionless con-
stants since the derivation of dimensionless constants was
one of the requirements for a final theory. The conclu-
sion of these considerations is that a FiT can not refer
to any physical constant at the most fundamental level.
Nevertheless it is supposed to provide objects with con-
stant properties which may then serve as references and
that provide natural units. This apparent paradox can
(only) be solved, if the constant properties of the objects
are constants of motion (COMs) as we shall explain in
Sec. V.
Only the existence of fundamental objects allows to
refer to fundamental quantities. This in turn allows to
express physical quantities by real (or even by natural)
numbers. The final theory became thinkable only by the
discovery of the sufficient number of physical constants,
e.g. by the invention of relativity and quantum mechan-
ics. Without the natural objective constants c and h̄, a
final theory can not be formulated due to the lack of ref-
erence quantities. Again: The values of these constants
are not meaningful, but their constancy is. This is the
modern understanding of physical units and constants:
A final theory contains no arbitrary units, because every
quantity can be expressed by the corresponding property
of some fundamental object of nature.
Dimensionless constants like the fine-structure con-
stant α ≈ 1/137 or the mass ratio of electron and pro-
tron me /mp ≈ 1/1836 represent proportions between the
properties of fundamental objects or processes. A FiT
should provide the mathematical arguments that allow to
derive these proportions and in case of success these con-
stants are effectively mathematical constants like π, which
represents an extremal proportion between circumference
and diameter.
D. Reference Objects and Constants of Motion
On the most fundamental level, a FiT may neither
refer to dimensional constants nor to dimensionless con-
stants, because the former are arbitrary and the latter
have to be derived within the FiT . But on the other
hand objects with constant properties are required as
references in order to allow for measurements. This ap-
parent contradiction can be resolved if we draw the fol-
lowing conclusions: Firstly, all quantities that belong to
the fundamental level, must be variables in the sense that
8
they have to vary at all times. Secondly, since constants
are required to enable for measurements, then the most
fundamental level obviously does not allow for a direct
measurement. However, there must be a one or several
functions of the fundamental variables that generate con-
stants of motion and therefore the reference constants
must be constants of motion. If all fundamental vari-
ables continuously vary, the constants of motion must
at least be of second order in terms of the fundamental
variables.
Constants of motion appear in all dynamical laws of
physics. The best known constants of motion are energy,
momentum and angular momentum and a FiT has to
provide a mechanism from which these COMs emerge.
From Noether’s theorems we know that constants of mo-
tion and invariance properties are closely connected. As
described in Ref.20,21 the suggested trivial principles al-
low to derive Hamiltons equations of motion (EQOM).
But the fundamental variables are not “coordinates of
mass points”, but purely abstract variables that can not
be directly measured. This is a logical implication of the
idea of a FiT and if we interpret fundamental variables
as (components) of quantum mechanics wave functions,
then we found a reason why the wave function itself can
not be directly measurered: Quantum mechanics has the
task to analyze the behavior of ensembles solely on the
basis of their observables. But the fundamental objects,
represented by the wave function, can not be directly
measured: There is no direct unit for the wave function,
because the (components of the) wave function vary at
all times.
A consequence of this argument is that truely funda-
mental quantities can never be subject of a direct mea-
surement. This is the trivial and evident reason, why we
can not measure (components of) ψ and claim that “the
first component of the wave function at position x is five
pumis” 6 , while there is no (principle) problem to obtain
the value of a magnetic field B(x) at some position in
units of Gauss or Tesla. Then, however, it follows that a
magnetic field value can not be regarded as a fundamen-
tal quantity: no directly measureable physical quantity
can be regarded as fundamental.
IV. WHAT COULD PROVIDE THE TRIVIAL EVIDENCE
OF A FINAL THEORY?
If we apply Goldbergs characterization of mathematics
to physics, then a FiT has to be a tautology. As men-
tioned in the introduction, we suspect that a FiT might
be formulated on the basis of what we do when we per-
form physical experiments, physical measurements, i.e.
what we have to assume unless our experiments and mea-
surements are meaningless. Here we might also find the
6 A pumi is a fictious unit for wave functions.

final answer to the question in what respect physics dif-
fers from pure mathematics.
Therefore the first step towards a FiT would be to
provide an inventory of the primary principles that are
trivially true in any world that allows for the formulation
of physical laws and to remove all presumptions that can
not be considered to be trivially fulfilled. Since substan-
tial presumptions can not be preconditions for the formu-
lation of a FiT , they should be critically reviewed. Es-
sentially this requires a description of the form of physics.
A. The Form of Physics: What Physicists Do
Mathematics is itself not a model of anything but a
logical framework, and also a FiT can not in it’s origin
be a model of anything. Hence it can not be directly
derived from the phenomena. If a FiT is possible, then
the description of the physical aspects of reality is based
on a set of formal principles and we will show that the
form of physics provides trivial principles that can be
used as a starting point.
Physics is a scientific discipline because it is based on
a well-defined method to test theories. The method of
physics is “objective” - but not because it describes “real”
objects or some kind of “objective reality”. It is objec-
tive because it is based on the comparison of objects with
objects. Measurements are comparisons - physicists com-
pare the size, weight, velocity, energy and other quantita-
tive properties with the size, weight, velocity and energy
of reference objects: the prototype meter, the kilogram
and the velocity of light are - or have been - used as ref-
erence. The branch of physics that provides the required
references, is called metrology. The definition of units
and the production of reference artifacts might appear to
be an inevitable but merely formal act. However, it is ex-
actly of the kind of inevitability which is required for the
formulation of a FiT . The inevitable aspects of physics
are those which in summary are the form or method of
physics. Since physical theories are tested by measure-
ments, the things that physisists do when they perform
measurements, is an essential part of the form of physics.
Einstein wrote about his theory of special relativity
that “It is striking that the theory (except for the four-
dimensional space) introduces two kinds of things, i.e.
(1) measuring rods and clocks, (2) all other things, e.g.,
the electromagnetic field, the material point, etc. This,
in a certain sense, is inconsistent; strictly speaking, mea-
suring rods and clocks should emerge as solutions of the
basic equations [...], not, as it were, as theoretically self-
sufficient entities.”11 . This confession should be surpriz-
ing and disconcerting to physicists and philosophers of
science, as it suggests that there might be missing piece,
a systematic incompleteness in the logic of objectivity of
the natural sciences. However (to the knowledge of the
author), the fundamental role that metrology plays with
respect to the method of physics is rarely mentioned, dis-
cussed or even noticed, despite the fact that the use of
9
objects as reference standards is undeniable the funda-
ment of objectivity in physics. A final theory, if consid-
ered possible at all, has to incorporate the trivial and
fundamental role played by measurement standards.
In the mentioned quote, Einstein describes what he
did when he formulated special relativity, but it is just
an example of what physicists always do, namely to in-
troduce references combined with certain invariance as-
sumptions. As they seem blatantly trivial, these assump-
tions are rarely spelled out explicitely; but to define rods
and clocks (e.g. some objects) as references, implies a
number of presumptions about the relevant properties of
these objects, namely that they are invariant. First of all
they are assumed to be constant in time, but the meter
rod is also assumed to be invariant with respect to its ori-
entation in space. Certainly we won’t suggest to doubt
these invariance principles, but we should be aware of
the implications of this method. Because the presumed
invariance principle leads, if violated, to the introduction
of fields: If the spin of an electron has no preferred direc-
tion, then this is phenomenologically identical to the ab-
sence of a magnetic field. And vice versa: The violation
of the isotropy of space is phenomenologically identical
with the presence of a magnetic field. Contrary to what
one might expect, but the presumed isotropy of space
does not require an experimental proof: Where we find
space to be anisotropic, we (have to) presume a reason,
namely a (gauge) field. Newton’s axiom of inertia obeys
the same logic: According to this axiom a body moves
with constant velocity on a straight line unless a force is
applied. Not only that, at the time of Newton, it was
technically impossible to provide experimental evidence
for the presumed straight motion, it is logically impossi-
ble to prove Newton’s premise as it is itself the definition
of (the absence of) a force. Therefore it is introduced as
an axiom. The underlying principle of this method is the
principle of (sufficient) reason. Science has to provide
reasons for distinctions, for asymmetries, for differences.
And if such reasons are not at hand, then we (have to)
assume symmetry, homogeneity, equivalence. This is not
a principle of nature, it is a principle of thought.
It is one of the revolutionary but also one of the most
questionable aspects of relativity, that it does not simply
continue to presume the invariance of rods and clocks,
but describes precise conditions of invariance and co-
variance and the corresponding transformation proper-
ties. And in doing so, it apparently breaks with a rule of
thought: Rulers are, as we explained above, by definition
constant. However, this holds only, if the we assume that
space-time is a fundamental notion, an assumption that
is, as we argued above, incompatible with the logic of a
FiT : it would deny the possibility to argue why space
should be three-dimensional. Whence there is no contra-
diction between relativity and the rules of thought, if a
FiT is possible. The FiT transforms the contradiction
into an obligation: The FiT has to formulate a criterion
that allows to characterize dimensionalities and to show
how spatial notions emerge.

Prior to the invention of relativity, physicists simply
presumed that clocks and rods are the same for all ob-
servers under all conditions. The theory of special rela-
tivity is the first theory that questioned this naive pre-
sumption. Now, after the experimental confirmation of
relativity, we believe to know that (perfect) rods are in-
variant in proper time and under rotations, but not un-
der Lorentz boosts. As rods are rarely used to measure
a length when they move with relativistic speeds, this
effect is irrelevant in all practical situations. It is dif-
ficult to imagine a situation in which the “length con-
traction” could have been discovered by the direct use of
rods for length measurements and accordingly it is diffi-
cult to confirm this “effect” by direct measurements. It
is a consequence of a formalism that has been confirmed
by a variety of other methods and predictions. Despite
this, the theory of special relativity has achieved the sta-
tus of an almost unquestionable corner-stone of physics.
Namely in accelerator and particle physics, it has reached
the same status of daily applied engineering knowledge as
electrodynamics and is usually regarded as part of clas-
sical physics. To the general public, relativity is usually
presented as a theory of space-time and rarely regarded
as an extension of metrology. We think that this might
be a mistake: Seen by light, Einstein’s work is full of
hints that relativity is intimately connected to problems
of measurement standards and relative calibration (clock
synchronization).
Our considerations result in the insight that it is not
the constancy of the length of a specific (inertial) ruler
that we have to presume, but we have to presume that
regardless of the technical problems to provide a really
constant ruler, it is in principle possible. This is part
of the form of physics: Despite all practical problems,
we presume that there is no general physical law that
prevents us in principle from manufactoring a constant
ruler or a perfect kilogram. As a final theory has to be
free of logical circles, the final theory can not presume
stability of the natural units and rulers and provide at
the same time an explanation for this stability.
A FiT can not be based on the assumption that a
specific property of a specific object is constant in time,
but it may be based on the more abstract assumption
that there are constants. It has to incorporate the basic
assumption of metrology, namely that nature provides
objects with constant properties that can be used as ref-
erences. It is then logically impossible to prove this as-
sumption within the theory. Any attempt to do so pro-
vides merely the proof of internal consistency: It would
just show that the theory allows to reproduce the as-
sumptions it is based upon.
B. Existence, Appearance and Philosophy
But how at all can it happen that preconceptions based
on empirical evidence fail, that matter is almost empty
space, that Newtons concept of absolute time and space
10
turned out to be wrong? Our preconceptions about space
and time are taken from “reality” as it appears to the
senses and as it is processed by our mind. But is it not
science (and specifically physics) that provides evidence
and confirmation for the existence of an objective real-
ity? Steven Weinberg dedicated a complete chapter of his
book to argue “against philosophy”1 . Nevertheless he ad-
mits on the second page of this chapter: “Physicists do
of course carry around with them a working philosophy.
For most of us, it is a rough-and-ready realism, a belief
in the objective reality of the ingredients of our scientific
theories.”
It is irritating that Weinberg argues against philos-
ophy, as his book appears to be more about philoso-
phy than about anything else. But as he explains, he
argues not against philosophy as such but against dog-
matism. He writes: “Although naive mechanism seems
safely dead, physics continues to be troubled by other
metaphysical presuppositions, particularly those having
to do with space and time.” and on the same page he
writes: “Even now, almost a century after the advent of
special relativity, some physicists still think that there
are things that can be said about space and time on the
basis of pure thought.” One might be tempted to remark
that the idea of a final theory, as Weinberg (pro-) poses
it, may as well suit the purpose of arguing in favour of
a theory derived from pure thought: The inevitability,
logical necessity and coherence that he expects from a
final theory, are properties that can otherwise only be
found in mathematics, that is: in a branch of science
that is based on pure thought. Again we arrive at an
irritating dichotomy. On the one hand, physics defines
itself as an experimental science, but on the other hand
it is the final objective of (theoretical) physics to find a
final theory which must be similar to pure mathematics,
in various aspects. However, the keyword in the above
quote is “metaphysical presupposition” and in this re-
spect the author fully agrees with Weinberg. A theory
should not (and a final theory may not) contain meta-
physical presuppositions. This is a key requirement that
we specified for a final theory: It may not contain any
substantial distinctions without sufficient reason. Specif-
ically we deny the possibility that a FiT could be based
on metaphysical assumptions. Within the context of a
FiT we have to refuse any assumption that is not based
on reason, any distinction that is not inevitable. And in
consequence this means that we refuse any substantial
claim.
It appears that Weinberg is aware of the fact that the
“rough-and-ready realism”, understood as a habit or a
tendency to be philosophically attached to the appear-
ance of reality, insofar as it is similar to “naive mechan-
ims”, is disproven by experimental evidence.
In his allegory of the cave, the greek philosopher Plato
adressed the question, whether appearance generates a
faithful image of the true form of reality: “Plato has
Socrates describe a gathering of people who have lived
chained to the wall of a cave all of their lives, facing

a blank wall. The people watch shadows projected on
the wall from things passing in front of a fire behind
them, and they begin to give names to these shadows.
The shadows are as close as the prisoners get to view-
ing reality.“9 . In Platos allegory the perception of real-
ity does not show something unreal or an illusion, but
something that is a projection or a derivation from the
true world 7 . Thus our preconceptions fail, because they
are based on notions derived from the appearance of the
world, namely (as Weinberg emphazises) on wrong pre-
conceptions about space and time. Our preconceptions
fail because they are based on “empirical evidence”, but
not on the kind of empirical evidence of a physics exper-
iment, but on the kind of naive “empirical evidence” of
common sense.
Weinberg wrote: “It is just that a knowledge of phi-
losophy does not seem to be of use to physicists” 8 . Here
we clearly disagree with Weinberg. First of all, we be-
lieve that some knowledge of philosophy is useful to any
human being, but more than that, we doubt that the use-
fulness of philosophy can be derived from the possession
of some kind of knowledge. Same as physics, philosophy
is mainly a proficiency. Because a final theory may not
be based on metaphysical assumptions, the proficiency to
recognize assumptions as metaphysical is required. Re-
gardless on how we arrive at a FiT - only if the theory
can be (re-) constructed on the basis of pure thought, it
can be said to have met his specifications. And philoso-
phy (even more than mathematics) is the domain of pure
thought.
Besides the major concepts of physics also the specific
(apparently relativistic) form of space and time including
Lorentz transformations and the dimensionality of space-
time must be the result of a final theory. In order to
obtain these notions as results from a final theory, the
theory has to be based on something else, e.g. not on
the presumption of a self-sufficient space-time model. It
has to include a part that describes the emergence (or
appearance) of spatial notions.
C. The Form of Dynamical Laws
The example of a physical law as described in Sec. III C
concerned some general relation between two physical
quantities like (for instance) energy and momentum.
Such relations allow to reduce the number of free vari-
ables by uncovering a fixed relation between quantities
(variables). If we discuss whether a specific process (or
7 Of course, Plato uses the allegory of the cave to claim that a
philosopher is someone who broke the chains and is therefore able
to see the world as it really is. This claim possibly contains a fair
amount of “marketing” for the case of philosophy. Otherwise the
famous (though slightly adulterated) quote “I know that I know
nothing” would likely not be found in Platos work10 .
8 Chapter 7, “against philosophy”.
11
nature as a whole) is deterministic (i.e. if it could be pre-
dicted by an algorithm), then we refer to a different kind
of physical law, namely to a dynamical law, to so-called
equations of motion (EQOM). EQOMs refer to a num-
ber of dynamical variables that can be expressed by the
components of a state vector ψ and have the archetypical
form
ψ̇ = f (ψ) , (11)
- where the dot indicates the time derivative - such that,
if the state of a system is represented by ψ at a certain
point in time, all future states of the system can be com-
puted by integrating Eq. 11. We do not consider partial
differential equations here, for two reasons: Firstly, we
said that a FiT is allowed to be based on the notion of
time, but not on space - and secondly, we expect that the
introduction of spatial notions can be delayed to a later
stage, as it is always possible to use a Fourier transfor-
mation to introduce spatial notions.
In classical physics and namely in classical statistical
mechanics, the state vector contains the coordinates and
momenta (Hamiltonian mechanics) or the coordinates
and velocities (in Lagrangian mechanics), respectively.
In the former case, the EQOMs are derived from a con-
servation law (usually the conservation of energy) and in
the latter case from a variational principle (the so-called
principle of least action). The EQOM depend on a single
variable t which can be interpreted as time, the primary
“quantity” that we accept as being undefinable. Time
is the (first) variable for which the question “why time”
makes no sense within a final physical theory: Already
the concept of physical variable and of measurement im-
plies that we refer to change, variation and constancy,
i.e. to those notions that are associated with time. And
we shall use the required distinction of variables and con-
stants to discuss the possible forms of Eq. 11.
V. A SURVEY OF (TRIVIAL) PRINCIPLES OF
PHYSICS
Our considerations allow to give a first, possibly in-
complete, inventory of formal first principles. The most
important being the principle of sufficient reason. Sir
Hamilton summarized this principle in the short formula:
“Infer nothing without a ground or reason.”8. The idea
of a FiT itself is based on this principle as we explained
in Sec. II.
The principle of reason (PSR) is the first principle that
trivially has to be respected in a FiT . The believe that
things are comprehensible is a precondition for science in
general, of course. But the principle includes more than
that. From the specifications of Weinberg, t’Hooft and
Susskind it follows that if a FiT is supposed to be in-
evitable, evidently true, complete and final, then a FiT
may not contain any arbitary distinction or selection, no
concept should be introduced without a clear and ev-
ident reason. The final theory does not have to pro-

vide causal explanations for individual phenomena, but
it should provide sufficient reasons for the general form
of the phenomena. Hence a FiT can not (and does not
have to) tell what time is, since we can not even formu-
late the idea of what physics is without presuming that
the concept of time is already known.
Secondly, as we explained before, the presumed con-
cepts of time and measurement inevitably lead to the
trivial confession that physical models contain variables
and constants. As we argued, for logical reasons, con-
stants may not appear at the fundamental level, neither
dimensional nor dimensionless. Despite this, the idea
of a measurement requires the appearance of true con-
stants for reference. We might call this requirement “the
principle of objectivity” (POO). It is the basis of the re-
lation between a physical theory and “objective reality”:
Physical equations refer to relations between properties
of objects and though this is rarely within the focus of
attention, these objects have to be physically made, if
the equations are to be confirmed. The production of
artifacts with constant properties must after all be pos-
sible in principle. If this is not the case, then we can
not directly measure this property. Even worse, strictly
speaking it could be questioned if such a property is ob-
jectively there.
The dichotomy that we may not directly introduce con-
stants while measurements are nevertheless based on the
availability of such constants, can be solved: dimensional
constants are logically connected to the laws of motion -
they appear in the form of constants of motion (COMs).
Within a FiT constants of motion are the only possi-
ble method to introduce dimensional constants. Either
we derive constants of motion from a law of motion, or
vice versa, but in the former case we have to presume
substantial laws, while in the latter case we just pre-
sume the trivial fact that constants must exist. The lat-
ter presumption is sufficienty trivial to serve as a first
principle, but we have a prize to pay: if the most fun-
damental constants are constants of motion that come
along with some law of motion of some set of fundamen-
tal variables ψ = (ψ1 , ψ2 , . . . , ψk ), then a reference for a
measurement of the components of ψ is logically impos-
sible. The fundamental level is a level of pure variables,
while the “level” of constants of motion is at least second
order in ψ. As we have shown in Ref.20,21 , the constants
of motion are the even moments of ψ. Since COMs are
required to obtain observables, then, within a FiT , the
fundamental level of reality can not be observed: The
final theory must distinguish between existence and ap-
pearance.
In Ref.20 we used an “ontology of time” as the basis
for the same reasoning. But as we have shown in this
essay, the suggested principle of variation (POV) can be
derived from the idea of a FiT , hence finally from the
principle of sufficient reason. A genuine ontology is not
required, as we derived the law, that all variables on some
fundamental level have to vary, from distinct formal con-
siderations.
12
Besides the distinction of variables from constants, we
have to distinguish quantity from structure. It is trivially
true that physical models contain variables that repre-
sent quantities. But structures have to be derived. A
quantity can be represented by a number and a unit. A
structure is given by an algebra. The complex plane for
instance is represented by an algebra: The unit imaginary
i can not be represented by a number and therefore we
(have to) introduce a symbol. Something similar holds
for the number π, namely the fact that we can not write
it down explicitely, suggests the use of a symbol. But
unlike π, which can by no means be completely written
down, we can represent the unit imaginary by a structure
that fully characterizes the algebra of the complex num-
bers, namely that of 2 × 2-matrices21 . The introduction
of a new symbol like i is equivalent to the introduction of
an algebraic structure. It is therefore reasonable to ask,
if the unit imaginary is essential for quantum mechanics
or not - and if it is, then why? The use of complex num-
bers in quantum mechanics can be questioned and hence
it is a substantial choice. Since a FiT may not presume
what it is supposed to derive, it is not legitimate to sim-
ply presume a prominent role for complex numbers in a
FiT ; the FiT has to provide a reason.
This prelimenary survey does not claim to be com-
plete. There are likely more trivial principles that belong
to the form of physics which we did not even mention.
One of the principles that we did not discuss is implied
in the possibility of a quantitative description: If there
are observable quantitative properties of objects, then
these properties must be additive: The length is of two
rods is twice the length of a single rod. And there are
certainly more principles of this kind. We summarized
these principles in the formula “a world that allows for a
physical description”, which is both, precise enough for
the purpose of this paper and vague enough to include
those principles that we did not discuss explicitely.
Let us finally emphasize that the describes principles
imply a strong selection: Whatever might be “out there”,
we can only measure those things that obey a law of mo-
tion which provides the necessary constants of motion as
reference standards. Everything else is noise. Scientists
should in general avoid to make claims about unmeasur-
able things. However, as we argued, with respect to a
final theory, we can not avoid to make at least one claim
about the wave-function, namely that its components are
not directly measurable. This is not unreasonable, but a
logical necessity.
VI. OBJECTIVITY AND CLASSICAL REALISM
As explained in the previous sections, it appears that,
given a final theory would be possible or even the only
logical consequence of the program of some specific ratio-
nalism, then this seriously questions that objectivity and
realism are compatible with each other; if the final the-
ory would be found and could be regarded as objectively

and evidently (trivially) correct, then physical appear-
ance and existence must be distinguished and classical
realism must be regarded as disproven. Or, if we choose
to believe that classical realism must be correct, then a
final theory is impossible and there must be something
unreasonable, i.e. some kind of significant but unrea-
sonable arbitariness in the world, something outside the
PSR, something that a FiT could not derive or explain.
Hence it appears that there is an intrinsic contradiction
between realism and rationalism. The rationalist would
expect that a FiT is possible - but then a “realistic”
worldview is logically impossible.
We defined the notion of objectivity of experimental
physics and we believe it is as clear as it is practically
and theoretically relevant, though it refers to the notion
of object, i.e. to something that is given only through
experience. Realism, the complementary notion of ob-
jectivity, on the other hand, is more difficult to specify,
also, because is appears in a number of variants. Often
realism and objectivity are used in combination as objec-
tive reality, some sort of common sense agreement about
what is meant by “the world”. Instead of making at-
tempts to give a theoretical definition of “realism”, let
us instead refer to a number of statements, that a strong
realist would likely support:
• The world is in a fundamental way as we perceive
it, i.e. composed of objects located in space and
time. Objects are made of matter.
• Every physical effect has an exclusively physical
cause: The physical world is closed. This includes:
• The only possible way to influence the physical
world is by physical action (i.e. through actions
of the body).
• All valid information about the world originates
from information that has been received physically,
i.e. with finite velocity v ≤ c via physical signals
through our senses.
It is noteworthy, that while the principle of objectiv-
ity defines (and therefore limits) the scope of empiri-
cal sciences, especially that of physics, realism defines
(and therefore limits) reality, most often to the notions of
space, time, matter and causality, or by suggesting lim-
its of possible information transfer. The former leaves
the question open if the objectively real is a subset of
the real, while the latter reduces the world to the objec-
tive by definition. Strange enough that these notions are
rarely distinguished by physicists. We do not intend to
discuss this is much detail, but we would like to empha-
size that the principle of objectivity defines a fundamen-
tal rule for quantitative empirical science that can not be
questioned on scientific grounds, while realism tries to es-
tablish a specific ideology that is based to a large degree
on unproveable metaphysical claims. One might summa-
rize this as follows: it is logically impossible to question
that physical laws are - within their limits of validity -
13
objectively true, unless one questions that a kilogramm,
a meter rod or a clock can be used as measurement stan-
dard. But classical realism reverses this logic and claims
that only the “objective” aspects of the world “really” ex-
ist. We think that this type of restrictive realism is not
only not required, but it builds a barrier between natural
and social sciences that is not helpful. Besides that, the
metaphysical assumptions of realism are not required to
obtain objective knowledge of nature.
Especially the closure claim of classical realism can be
considered problematic, especially in the context of the
questions of free will and the nature of consciousness.
VII. SOME FORMAL CONSIDERATIONS
While we measure the weight of an object, by compar-
ing it to reference weights, we do not bother too much
about the number system used. This might be depre-
cated by pure mathematicians, but for most experimental
physicist (and likely for most engineers), there are only
two types of numbers required: integers and reals. The
question, whether it is logically necessary to use the reals
instead of the rational numbers, can be pragmatically an-
swered insofar as there is no immediate reason to restrict
the number system to the rationals, despite the fact that
we can never obtain a non-rational measurement result
in measurements of limited precision. Though we can not
measure π with infinite precision, we are still be able to
find ways to measure π to any required precision by the
use a sufficiently large circles. Insofar the experimental
determination of π via measurement is in principle not
different from its computation: No one can compute (or
write down or store on a computer disc) the number π
with infinite precision; but it can be done with any re-
quired precision: there is no logical limit to the precision
with which we can either compute or measure π.
Sometimes it is argued that the reals are - besides the
complex numbers and the quaternions - just the simplest
field over which one can define division rings and that
this fact suggests that the most general mathematical
objects (i.e. the quaternions) should be used on the fun-
damental level of a final physical theory. Despite the fact,
that quaternions might play a prominent role in some ar-
eas of physics, we reject the general argument. Both the
complex numbers as well as the quaternions, can be rep-
resented by real valued structures, namely the complex
numbers by real 2 × 2 and the quaternions by real 4 × 4
matrices. Therefore they are not more fundamental than
real matrices, but less, as they refer to specific matrix
algebras. Instead of presuming a fundamental role for
a specific algebraic structure, a FiT should derive and
explain this structure and prove its fundamentality. The
intransparent mixture of quantity and structure however
may lead to confusion about the meaning and role of
complex numbers in quantum mechanics and physics in
general. This is indicated by the discussion about the
meaning of complex numbers in QM that is ongoing for

decades22–34 .
We do not recognize the progress in mixing quantity
and structure on the fundamental level. Rather, if com-
plex numbers are a preferable, this should be a result
of the algebraic properties of the FiT . This fact is re-
flected by the architecture of number systems in com-
puters, computers which are able to perform simulations
of quantum systems, i.e. integers and reals. A com-
puter is usually understood as purely based on binary
digits. Though this is trivially true for contemporary
digital computers, the existence of analogue computers
and the fact that the personal computer had originally
two processing units, the central processing unit (CPU)
and the floating point unit (FPU), can be regarded as an
argument against the fundamentality of the bit: Even, if
the memory of modern computers is by design based on
bits, it is undeniable that we can not use it in a general
scientifically meaningful way without introducing a repre-
sentation of floating point numbers. This representation
is based on bits for purely practical reasons. Though any
given coding scheme of floating point numbers allows only
for a limited precision, we are able to construct coding
schemes for any required precision.
A. Structure Preservation
Any finite physical system can be described by a (possi-
bly infinite) set of varying real fundamental quantities as
a function of time ψ(t). But all observable (measurable)
properties are of even (at least second) order in these fun-
damental variables. The variables themselves can not be
directly measured due to the lack of constant references.
References for measurements are constants of motion and
they are at least of second order in ψ. Once we presume
some (arbitrary) constant of motion H(ψ), then further
constants of motion emerge in the description of this fun-
damental “objects”20,21 , namely the traces of powers of
the autocorrelation matrix (matrix of second moments),
multiplied by the symplectic unit matrix. When the gen-
eral constant of motion H(ψ) can be written as a (multi-
variate) Taylor series, then the leading order terms auto-
matically generate structure preservation. The presumed
constancy of H(ψ) generates a self-identity of “objects”
via the identity of motional patterns. No other identity
is required and possible. As we argued in Ref.20,21 , the
most general framework for structure preserving motion
is symplectic motion and it is a logical consequence of the
constancy of H 9 . Furthermore it has been shown that
the most simple, fundamental, and non-trivial algebra
that describes the general properties of symplectic mo-
tion, is given by the real Dirac (Clifford-) algebra Cl(3, 1)
which implies that the fundamental set of observables
9 It can be shown that unitary motion is a special case of sym-
plectic motion35
14
suggests the pattern of Minkowski space-time. There are
10 sympletic generators within this Lie-algebra that re-
quire an interpretation: a 3 + 1 dimensional vector and a
six-component bi-vector. The transformation properties
of these components suggest the interpretation in terms
of energy, momentum, electric- and magnetic field20,21
and the equations of motion of the autocorrelation matrix
allow for a derivation of the Lorentz force. However, the
primary and fundamental variables are not space-time co-
ordinates any more. Instead the fundamental variables
are the (abstract) components of ψ, which can best be
described as points in an abstract phase space. These
ideas thus reverse the role between “real” space and
phase space: The 3 + 1-dimensional “energy-momentum
space” is an effective space that emerges on the basis
of a fundamental 4-dimensional phase space described
in proper time and the concept of space-time emerges
through Fourier transformation. Effectively this Fourier
transformation is similar to the characteristic function of
probability theory, but its argument is not the probabil-
ity density function (PDF) ρ, but the product of ψ and
√
ρ. The 6-dimensional phase space of a “mass point”
emerges from the equations of motion of the second mo-
ments and is symplectic only in a low-energy and low-field
approximation.
Hence it seems possible to derive the concepts of mass
and spin (the two invarients of motion of a 4-spinor), of
energy-momentum 4-vector, of the electromagnetic field
and the transformation properties of the Lorentz group
(rotations and boosts) on the basis of trivial principles.
This is a first “proof of principle” that a FiT might in-
deed be possible - as a tautological theory.
VIII. SUMMARY
In this paper we picked up the idea of a final theory of
physics. First we formulated the requirements of a final
theory. These seem to be extreme in view of the severe
difficulties of modern physics to unify gravity and quan-
tum mechanics. But regardless of the current status of
theoretical physics, the idea of a final theory has its roots
in the principle of sufficient reason (PSR) and hence es-
sentially in rationalism. Specifically the idea that a final
theory would be inevitable and that it would carry the
evidence for its truth in itself seems to suggest that -
following Goldberg’s characterization of mathematics - a
final theory of physics must be a tautology. This implies
that all statements of the theory are predetermined in
its premises such that it is possible to derive the content
of physics logically from the premises, but also to elabo-
rate to a higher degree of precision what a final theory is
able to tell us about the world, and even more, what it is
not able to tell. The absence of metaphysical presuppo-
sitions is a logical necessity of the finality of a FiT - but
it does not imply the absence of “metaphysical” conclu-
sions. The final theory as we scetched it, might tell a lot
about reality, but likely not what some of us expected to

find. It will not tell us, why there is something instead of
nothing. It will not even tell us, what type of substance
matter “really” is. The tautological FiT - and we claim
that no other FiT is possible - is, well, tautological.
Besides the PSR, which provides the rationale for a fi-
nal theory, and is therefore obligatory, we discussed some
premises for a FiT that may in summary be called the
form of physics. We presented a first primitive analysis of
the conjectured possibility of a FiT . We identified three
types of physical constants, namely dimensionless con-
stants, dimensional conversion factors and natural units.
We explained why the only possible dimensional physical
constant within a FiT is the natural unit and that it can
only emerge from the theory as a constant of motion. We
argued that this result logically implies that the funda-
mental level of a FiT is not directly measurable - not
because the fundamental level is less “real”, but because
there is an unavoidable lack of reference.
REFERENCES
1 Steven
Weinberg; “Dreams of a final theory”, Vintage Books
(1993), ISBN 9780099223917.
2 G. t’Hooft, L. Susskind, E. Witten, M. Fukugita, L. Randall,
L. Smolin, J. Stachel, C. Rovelli, G. Ellis, S. Weinberg and R.
Penrose; Nature Vol. 433, 20th Jan. 2005, pp. 257-259.
3 A. Einstein in: The Living Age, Vol. 304 (3 Jan 1920), 41-43.
4 Wikipedia
on “Reality” (Jan. 2017):
https://en.wikipedia.org/wiki/Reality.
5 Max Tegmark: “The Mathematical Universe” (2007);
arXiv:0704.0646.
6 W. Smilga “Reverse Engineering Approach to Quantum Electro-
dynamics”; J. of Mod. Phys. Vol.4 (2013), 561-571.
7 “Principle of Sufficient Reason”, Stanford Encyclopedia of Phi-
losophy; https://plato.stanford.edu/entries/sufficient-reason/
8 Sir William Hamilton “Lectures On Logic Vol. 1”, Boston: Gould
(1859), https://archive.org/details/lecturesonmetaph00hamiuoft
9 Wikipedia on “Allegory of the Cave”,
https://en.wikipedia.org/wiki/Allegory of the Cave, dated
August 14th, 2016. Original: Plato Republic, 514a.
10 Wikipedia on “Sokrates”, https://en.wikipedia.org/wiki/Socrates,
original: Plato Apologeia, 21d: “Denn es mag wohl eben keiner
von uns beiden etwas Tüchtiges oder Sonderliches wissen; allein
dieser meint etwas zu wissen, obwohl er nicht weiss, ich aber,
wie ich eben nicht weiss, so meine ich es auch nicht.”.
11 A. Einstein, Autobiographical Notes, Trans. by Paul Arthur
Schilpp, Open Court Publ. 1979.
12 Antonis Modinos; “From Aristotle to Schrödinger”, Springer
(Cham, Heidelberg, New York, Dortrecht, London) 2014.
13 Dino Boccaletti, “Galileo and the Equations of Motion” Springer
(Cham, Heidelberg, New York, Dortrecht, London) 2016.
15
14 Wikipedia; https://en.wikipedia.org/wiki/Physical constant,
(5th July 2016).
15 P.A.M. Dirac “The Large Numbers Hypothesis” in: The Physi-
cist’s Conception of nature; Edt. Jagdish Mehra (1973), D. Reidel
Publishing Company Dordrecht-Holland / Boston-U.S.A..
16 Hsu, L. and Hsu, J. P. ”The physical basis of natural units and
truly fundamental constants” (2012). Eur. Phys. J. Plus, 127: 11.
17 John P. Ralston “Quantum Theory without Planck’s Constant”,
arXiv:1203.5557, (2012).
18 Stanley Goldberg “Understanding Relativity”, Birkhäuser Basel
Boston Stuttgart (1984).
19 C. Baumgarten “Use of real Dirac matrices in coupled linear
optics”; Phys. Rev. ST Accel. Beams. 14, 114002 (2011).
20 C. Baumgarten; “Relativity and (Quantum-) Electrodynamics
from (Onto-) Logic of Time”; in Ruth E Kastner, Jasmina
Jeknic-Dugic and George Jaroszkiewicz (Edt.) “Quantum Struc-
tural Studies”, World Scientific (2017), ISBN: 978-1-78634-140-2.
Preprint arXiv:1409.5338v5 (2014/2015).
21 C. Baumgarten; “Old Game, New Rules: Rethinking the Form
of Physics”, Symmetry 2016, 8(5), 30; doi:10.3390/sym8050030.
22 F. Strocchi, “Complex Coordinates and Quantum Mechanics”,
Rev. Mod. Phys. 38, No. 1 (1966), pp. 36-40.
23 D. Hestenes “Vectors, Spinors and Complex Numbers in Classical
and Quantum Physics”, Am. J. Phys. Vol. 39/9 (1971), 1013-
1027.
24 W.E. Baylis, J. Huschlit and Jiansu Wei “Why i?” Am. J. Phys.
Vol. 60 (1992), 788-795.
25 Philip Goyal, Kevin H. Knuth and John Skilling “Origin of com-
plex quantum amplitudes and Feynman’s rules”, Phys. Rev. A
81, 022109 (2010).
26 Julian B. Barbour, “Time and complex numbers in canonical
quantum gravity”, Phys. Rev. D 47, No. 12 (1993), pp. 5422-
5429.
27 J.L. Rosales “Remarks on the issue of time and complex numbers
in canonical quantum gravity”, Phys. Rev. D 54 No. 6 (1996),
pp. 4185-4188.
28 Matthew McKague, Michele Mosca and Nicolas Gisin “Simulat-
ing Quantum Systems Using Real Hilbert Spaces”, Phys. Rev.
Lett. 102, 020505 (2009).
29 Felix Lev “Why is Quantum Physics Based on Complex Num-
bers?”; Finite Fields and Their Applications, Vol. 12, 336-355
(2006), preprint: arXiv:hep-th/0309003.
30 Charis Anastopoulos “Quantum vs stochastic processes and the
role of complex numbers”; Int.J.Theor.Phys.42:1229-1256,2003,
preprint: arXiv:gr-qc/0208031.
31 J.A. Miszczak “States and channels in quantum mechanics with-
out complex numbers”, arXiv:1603.04787.
32 C.P. Kwong “The mystery of square root of minus one in quan-
tum mechanics, and its demystification”, arXiv:0912.3996.
33 James M. Chappell, Azhar Iqbal, Derek Abbott “Geo-
metric Algebra: A natural representation of three-space”,
arXiv:1101.3619v4.
34 G. W. Gibbons “The emergent nature of time and the complex
numbers in quantum cosmology”, arXiv:1111.0457v1.
35 J.P. Ralston: “PT and CPT quantum mechanics embedded in
symplectic quantum mechanics”; J. Phys. A Math. Theor. Vol.
40 (2007), pp. 9883-9904.